Cargando…

A new method for identifying industrial clustering using the standard deviational ellipse

Industrial agglomeration has attracted extensive attention from economists and geographers, yet it is still a challenge to identify the multi-agglomeration spatial structure and degree of industrial agglomeration in continuous space—there is still a lack of a more targeted industrial clustering meth...

Descripción completa

Detalles Bibliográficos
Autores principales: Zhao, Ziwei, Zhao, Zuoquan, Zhang, Pei
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Nature Publishing Group UK 2023
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9834335/
https://www.ncbi.nlm.nih.gov/pubmed/36631559
http://dx.doi.org/10.1038/s41598-023-27655-8
Descripción
Sumario:Industrial agglomeration has attracted extensive attention from economists and geographers, yet it is still a challenge to identify the multi-agglomeration spatial structure and degree of industrial agglomeration in continuous space—there is still a lack of a more targeted industrial clustering method. The clustering method and the standard deviational ellipse (simply, ellipse) model have advantages in identifying the spatial structure and representing spatial information respectively. On this basis, we propose an ellipse-based approach to identifying industrial clusters. Our ellipse-based approach rests upon group nearest neighbor using the group-based nearest neighbor (GNN) ordering and spatial compactness matrix, where a number of point sequences with varying lengths, generated under the GNN ordering, are characterized by an ellipse and the elliptical parameters of these point sequences formulate the values and structure of the compactness matrix. Clustering is reformulated to identify ellipses with a specified parameter among a number of potential candidate ellipses, with significant changes (especially in the area) used as the cutoff criterion for determining the clusters’ border point. Our approach is illustrated in the location pattern of firms in Shanghai City, China in comparison with four well-known clustering methods. With the combination of elliptical parameters and spatial compactness, our approach may bring a new analytical ground for future industrial clustering research.